How gut bacteria survive viral attack
ISB researchers use single-cell sequencing to reveal hidden defenses against bacteriophages — and introduce a powerful new way to study phage-microbe interactions.
In the microscopic world of the human gut, bacteria are constantly under attack.
Viruses called bacteriophages — or simply phages — infect bacterial cells, hijack their machinery, and burst them open to release new viruses. These microscopic battles shape the microbial communities that influence digestion, immunity, and disease. But scientists have long struggled to watch those battles unfold.
A new study published in the journal Nature Communications from researchers at the Institute for Systems Biology (ISB) offers a way to see them in unprecedented detail. Writing in Nature Communications, the team used bacterial single-cell RNA sequencing to track how thousands of individual cells of the gut bacterium Bacteroides fragilis responded to infection by a lytic phage — a virus that replicates inside bacteria and ultimately destroys them.
“Phages are incredibly important in microbial ecosystems, but they’ve been difficult to study at the level of individual infection events,” said Dr. Anna Kuchina, assistant professor at ISB and senior co-corresponding author of the study. “Most traditional methods average signals across millions of infected cells, which hides the variation that often determines whether infection succeeds or fails.”
To uncover that hidden variation, the researchers analyzed more than 50,000 individual bacterial cells. They infected cultures of B. fragilis with a newly isolated phage from King County wastewater and then captured a single snapshot of gene activity across thousands of cells at once.
That snapshot contained cells at many different stages of infection — allowing the researchers to reconstruct how infection progressed.
Some cells were clearly infected, packed with viral transcripts, and on the path to destruction. Others remained untouched.
By comparing these cells, the researchers could see both how the virus commandeered the host’s machinery and why certain bacteria escaped infection.
One way to picture the process is as a battlefield: phages raining down like arrows, while some bacteria happen to carry shields.
In this case, those shields were not emergency defenses activated after an attack. Instead, there were pre-existing differences within the bacterial population.
Some cells expressed surface features that made infection less likely. Others carried traits that left them more vulnerable. These differences created small subpopulations of bacteria that could survive the viral assault and regrow — without new mutations.
“That was one of the most surprising findings,” said Anika Gupta, co-first author of the study and a graduate student in the Kuchina Lab. “Resistance didn’t necessarily arise after the infection. In many cases, the protective traits were already present in a subset of the population.”
The study also highlights a new tool for studying phage biology. The researchers used microSPLiT, a single-cell RNA sequencing technology developed by Kuchina, to capture gene activity in individual bacterial cells. Instead of averaging signals across entire bacterial populations, the method reveals what individual cells are doing during infection.
Traditionally, studying phages requires isolating both the virus and its bacterial host and then observing their interaction in the lab. But many phages cannot easily be isolated, leaving large parts of the viral world unexplored.
By contrast, the new approach can reconstruct infection dynamics from a single snapshot of thousands of cells.
“This study represents a new application of bacterial single-cell RNA sequencing to bacteriophage interactions,” said co-first author Dr. Dmitry Sutormin, a postdoctoral fellow in the Kuchina Lab. “It opens the door to studying phages much more broadly in complex microbiomes.”
Understanding those interactions could have important implications.
Phages play a major role in shaping microbial ecosystems, including those in the human gut. They are also being explored as potential therapeutic tools to target bacterial pathogens, especially as antibiotic resistance becomes a growing global challenge.
But bacteria can evolve ways to evade phage infection. By revealing the traits that allow some cells to survive, the new research offers insights that could help scientists better predict — and potentially control — those outcomes.
More broadly, the work shows how studying microbes one cell at a time can reveal biological dynamics that remain invisible when averaged across entire populations.
“It’s a reminder that microbial communities are not uniform,” said Kuchina. “Even within a single species, individual cells can behave very differently — and those differences can determine the outcome of an infection.”
This research was a collaborative project that involved ISB’s Kuchina Lab, Neelendu Dey’s lab at Fred Hutchinson Cancer Center, and others. Read the Nature Communications paper here.
– Hi, I am Anna Kuchina. I’m an assistant professor at Institute for Systems Biology where we study bacterial physiology and bacterial behavior one cell at a time. And so today we’ll be discussing the paper that we just published in Nature Communications, where we looked at the interactions of a gut microbe, Bacteroides fragilis, with the bacteriophage, or the virus that infects it and kills it, that we isolated from Seattle municipal wastewater. And so in this paper we study this pathogen because it’s normally a commensal gut microbe that is generally harmless, but sometimes it turns into a pathogen, for example, when the intestinal wall is breached and it escapes into the bloodstream and causes infections. And so we were interested in how this bacteria interacts with the phage, in the process of the phage infecting them and killing them using our technology that we developed for studying gene expression in tens of thousands of microbial cells at a time. The lead authors of this study are Anika Gupta, a molecular engineering graduate student at University of Washington, and Dmitry Sutormin, a postdoc in my group. This is also a collaborative study with Neil Dey at Fred Hutchinson Cancer Center and his postdoc, Norma Morella, who is also a co-lead author of the study. So Anika, what can you tell us that we found when we looked at the individual cells of Bacteroides fragilis, this bacterium when it was challenged with the phage?
– Yeah, so when we looked at individual B. fragilis cells exposed to phage, we could really see these infected cells, in which the virus was replicating, and we were also able to computationally reconstruct the different stages of phage infection and see really how are these bacterial and viral expression states changing over time. And interestingly, we were also able to detect a fraction of cells that were not actually infected by phage. And that kind of tells us that within this phage treated population, not all cells are equally susceptible to phage.
– Well, this is surprising, right? Because all of the cells are supposed to be genetically the same, because they’re descendants of the same cell that we grew in the lab. But when we challenged it with phage, then we saw that a fraction of cells actually remained uninfected. So Dmitry, maybe you can comment on, how did the cells protect themselves? What did we find?
– Yeah, indeed. And one of the strongest signals we saw in our data when we compared infected and uninfected cells, we observed a specific selection of certain genes related to the capsular biosynthesis. And it turned out that some of these capsules, they are protective, so they allow bacteria to survive during the phage infection. But other capsules are sensitive. So almost all the cells having these sensitive capsules, they were heavily infected with a phage, and they’re going to die because of this infection. But that’s not the only thing we observed, actually. And interestingly, it turned out that the Bacteroides fragilis, it has a very plastic genome. It means that it is exposed to multiple chromosomal rare regiment. A lot of recombination is happening here and other genetic losses, they are also contributing to the phage infection or to phage protection. Among the genes we outlined a couple very novel and interesting candidates, for example, several clusters related to the fimbriae biosynthesis. The fimbriae are bacterial appendages built of proteins, they may be used for adhesion, to biofilm formation. So it’s work in progress. We didn’t really understand now what’s the mechanical basis for that, but we just picked up these candidates and observe that they contribute into the protection of the population.
– Thanks Dmitry, and so it’s also well known, isn’t it, that bacteria actually have immunity systems, so to speak, or defense systems that specifically protect them from viruses such as phages. So maybe, Anika, you can comment how, what did we find about the involvement of these immune systems in protecting the cells from the phage?
– Yeah, so these bacterial defense systems also did seem to play a role in protection against phage. We observed that in individual cells, the expression of defense systems was very variable. And in certain cases, with some defense systems, we observed that they were able to increase phage protection when expressed alongside one of those protective capsular polysaccharides.
– Wow, that’s fascinating. Thank you. So I guess to wrap up, maybe Dmitry you can comment what are the implications of the study and what are the next steps for us to applying our high throughput technologies for studying bacteriophage interactions?
– Thank you, Anna. So for me, this is first of all a very nice biological story that highlights how adaptive and plastic the bacterial systems are. We used to think that bacteria and viruses, they’re kind of robots that has some hardwired genetic circuits, and they respond to specific stimuli in a specific way. But this study shows that there’s a highly multi-layered system and bacteria can respond to the viral infection in multiple ways. And all these layers they interact with each other. There is a capsular polysaccharide layer, there is a fimbrial layer, that is defense system layers. They all interacted. And I think that’s not the end of the story, actually, a small cliffhanger here. We recently touched like another mechanism here. So stay tuned and maybe we’ll see another player in this playground. Speaking of a broader perspective, like how it is related to the phage therapy, because the phage therapy is an evolving field and it’s a very nice alternative to the antibiotics, because it’s more precise and it can handle pathogens which are not susceptible to antibiotics. But imagine if you have a pathogen of a similar plasticity, of a similar adaptability. That’s a huge challenge for the phage therapy because bacteria may have multiple ways to escape this type of treatment. And our single cell studies, they’re paving the way to elucidate all these multiple mechanisms in nature and how this mechanism are interacting to each other. So potentially using single cell transcriptomics, we can find a specific types of treatment, a specific composition of a phage cocktail which will cover the entire adaptive landscape of bacteria and which didn’t give them a chance to escape from our treatment.
– Thanks Dmitry. This is fascinating again, and the phage therapy angle is really an unexpected angle for us, because we are mostly interested in very basic questions, such as what are the potential implications of bacterial variability, in terms of expressing different genes and different parts of the bacterial populations to bacterial lifestyles. But now given that we’ve used our technology, single cell transcriptomics, to profile interactions between the bacterium and the phage at a single time point, but we are able to, using this snapshot, reconstruct the full cycle of bacterial infection and also find the cells that were protected, again from a single snapshot, gives us now the opportunity to extend our studies to studying these interactions in the context of the microbiome itself, right? Without potentially culturing or isolating the phage and the host pair. And so this potential that we sort of unlocked with this study and we demonstrate how we use our technology, microSPLiT, to do this, this potentially can open the door in profiling and solving one of the big challenges of studying microbiome phage interactions, because it is actually very hard, a very hard challenge to isolate the bacterium, the host bacterium and the phage sometimes, right, to study their interactions in vitro in the lab. But if we can just take the microbiome sample and take it as is without any transformations or culturing and then apply our technology, we will be able to similarly, as described in this paper, computationally reconstruct these processes. Thanks and congratulations on the paper. Please check it out, it’ll be open access in Nature Communications.
